Node Selection Method for Split Attribute in C4.5 Algorithm Using the Coefficient of Determination Values for Multivariate Data Set

Abstract

The split attribute in the decision tree algorithm, especially C4.5, has an important influence in producing a decision tree performance that has high predictive performance. This study aims to perform an attribute split in the C4.5 algorithm using the value of the termination coefficient (R2/R Square) which is combined with the aim of increasing the performance of the model performance produced by the C4.5 algorithm itself. The data used in this research are public datasets and private datasets. This study combines the C4.5 algorithm developed by Quinlan. The results in this study indicate that the use of the R2 value in the C4.5 algorithm has good performance in terms of accuracy and recall because three of the four datasets used have a higher value than the C4.5 algorithm without R2. Whereas in the aspect of precision, it has quite good performance because only two datasets have a higher value than the performance results of the algorithm without R2.

Implementation of C4.5 and K-Nearest Neighbor to Predict Palm Oil Fruit Production on Local Plantations

“…The fleshy part of the oil palm fruit will produce crude oil which is then processed into raw material for cooking oil. The remaining processing is used as animal feed and mixed ingredients which are fermented into compost (Mehraban et al, 2022;Muhsi et al, 2023;Tsujino et al, 2016).…”